Bharati Vidyapeeth College of Engineering

A paradigm shift is occurring in the manufacturing sector globally due to the digitization of operations; however, studies specific to the maintenance aspect are scarce. The research presents challenges particular to the implementation of smart maintenance in Indian manufacturing SMEs, which is unique and helps bridge the gap. This research aims to identify and analyze the hierarchical interrelationship among critical barriers to smart maintenance implementation in Indian manufacturing Small to Medium-sized Enterprises (SMEs). This study involves a two-phase approach. In the first phase, the critical barriers of smart maintenance in Indian manufacturing SMEs are identified. A literature review, along with the opinions of Industrial and academic experts, resulted in the finalization of the barriers. In the second phase, the Interpretive Structural Modeling (ISM) and Cross-Impact Matrix Multiplication Applied to Classification (MICMAC) methodologies were employed to analyze the criticality of the barriers. The driving dependence power was computed for each factor involved, which was then given as an input into the MICMAC analysis to cluster the barriers into autonomous, dependent, linkage, and independent barriers. Smart maintenance is gaining traction worldwide in the manufacturing sector, driven by the emergence and popularity of Industry 4.0. The study identifies the critical hindrances to the adoption of innovative practices specific to the maintenance function of Indian SMEs. ‘Limited Interoperability of Digital Systems’ and ‘Poor Data Quality and Fragmented Data Infrastructure’ emerged as the critical barriers with high driving power. At the same time, ‘Management Resistance and Low Strategic Priority’ and ‘Lack of Cross-functional Collaboration’ appear at the top of the ISM model with high dependence power. The overall approach of smart maintenance, as presented in this research, challenges identification and structured modelling, and is expected to help managers and policymakers concentrate their efforts based on the criticality of barriers. In turn, this will facilitate a smooth transition to smart technologies in the maintenance function of Indian manufacturing SMEs.

As electric vehicles (EVs) gain popularity, enhancing battery charging systems and health management is critical for extending battery life and driving range. This research proposes an improved AC-DC power-factor-corrected resonant converter integrated with a cascaded modular multilevel three-level inverter using stacked SiC MOSFETs, an isolated LLC converter, and a front-end rectifier to optimize EV charging efficiency. To accurately predict battery State of Health (SOH), a long short-term memory (LSTM) neural network combined with an attention mechanism is employed. The LSTM model captures temporal dependencies in battery data, while the attention mechanism emphasizes the most relevant features over time, improving prediction accuracy. The model is trained on a dataset with six key features, including current, voltage, and temperature, and optimized using a Differential Evolution algorithm for hyperparameter tuning. Simulation results demonstrate a 15

The injection molding is under continuous pressure to reduce waste and enhance quality without affecting the safety. Although Industry 5.0 advocates human-centric and sustainable production, the transition to the so-called Zero-Defect production is challenging since polymers can easily act unpredictably when subjected to heat and pressure. This paper presents a Responsible, Risk-Aware, and Regulated AI (RRRAI) model. We integrated a custom-made automated part separator with a high-fidelity Digital Twin to intercept defects in real-time. We did not apply deep learning models that are black boxes and cannot be trusted by the operators, but rather the Decision Tree Regression. This is associated with injection parameters namely temperature, pressure, and cycle time directly correlated with defects and is transparent to the system. An experiment on a FANUC ROBOSHOT 100 in the field revealed that our system could cut down on daily defects out of the average of 765.8 defects to 299.8 defects $(\mathrm{p}<0.001)$. The model had the prediction accuracy of 85 percent. Importantly, we ensured that the system is risk-conscious: when the AI is not certain (between 10-80 percent) then it requests a human inspection, as opposed to making guesses. The strategy is consistent with the NIST AI Risk Management Framework, which confirms that it is possible to be responsible and automate quality control.

The growing effects of climatic variability on crop production have heightened the necessity of the sophisticated predictive tools that can provide precise and interpretable crop yield predictions. This paper examines the use of the Explainable GeoAI models to forecast crop yields in South Asia, a place that has been extremely susceptible to climate shocks including unpredictable rainfall, climatic extremes, and droughts. The combination of geospatial information, inputted remote sensing data and climatic variables with the machine learning models enables the proposed framework to increase the predictive power of the models whilst making the models understandable using explainability methods. Approaches like SHAP (Shapley Additive Explanations) and feature importance mapping are used to explain the role of environmental and climatic factors on the yield results. This research also builds on multi-source data sets, such as satellite images, soil properties and past yield data to create an effective predictive model that is sensitive to the local agricultural environment. The findings have shown that Explainable GeoAI models are not only more effective at making predictions, but also offer useful information to the policymakers, farmers, and agricultural planners. The results contribute to the necessity of integrating the interpretable artificial intelligence with spatial intelligence to aid climate resilient agriculture and decision-making. This study will provide a contribution to the sustainable agriculture system in that it helps to bridge the gap between high-performance AI systems and their application in areas with the climate sensitivity.

Tailpipe emissions from CI engines continue to be a significant environmental and health issue. Biodiesel will lower greenhouse gases but will often increase NOx and decrease efficiency slightly. Oxygenated additives enhance atomization and efficiency but can cause additional NOx, whereas conventional EGR reduces NOx at the expense of efficiency. Reformed EGR (rEGR) provides a superior trade-off by producing hydrogen rich reformate from unburnt hydrocarbons. This work evaluates B20 Chlorella vulgaris biodiesel with 10% Butyl Methyl Ether (BME) under rEGR rates of 5 - 20%. The B20-BME mixture with 10% rEGR has an 8.7% increase in brake thermal efficiency and a 22% reduction in fuel consumption. Smoke, NOx, CO, and HC decrease by 34.2%, 16.1%, 50% and 39.5% respectively. Response Surface Methodology is used for confirmation of model validity, adjusted-predicted R2 deviation <0.2. The optimized BME-rEGR combination achieves higher efficiency with consistently lower emissions, including NOx.